Re: Eric Rescorla's No Objection on draft-ietf-httpbis-early-hints-04: (with COMMENT)

On Wed, Aug 02, 2017 at 03:03:53PM +1000, Martin Thomson wrote:
> On 2 August 2017 at 13:15, Willy Tarreau <w@1wt.eu> wrote:
> > On Wed, Aug 02, 2017 at 10:31:55AM +1000, Martin Thomson wrote:
> >> On 2 August 2017 at 07:50, Eric Rescorla <ekr@rtfm.com> wrote:
> >> > I don't understand this text:
> >> > "   HTTP/2 ([RFC7540]) server push can be used as a solution to this
> >> >    issue, but has its own limitations.  The responses that can be pushed
> >> >    using HTTP/2 are limited to those belonging to the same origin.
> >> > "
> >> >
> >> > Isn't this also a limitation of 103?
> >>
> >> Yes.
> >
> > Hmmm no, these are hints containing Link header fields, so they can
> > reference any object including external ones just as if they were
> > delivered in the final response or in the HTML header.
> 
> Oh, the claim is about the resources that can be referenced, which
> could be anywhere, for which I agree.  The comparison with server push
> is questionable given that server push can deliver resources in
> addition to the identifier.

Indeed. Maybe it should be presented like a complement and not an
alternative. Some sites might want to continue to deliver very small
resources using PUSH and larger ones using 103 for example (so that
clients can cache them and in order to avoid wasting bandwidth).

> Part of my confusion was that "this issue" has become a little muddy
> after 4 paragraphs.  I would instead say:
> 
> """
> HTTP/2 server push [ref] can accelerate the delivery of resources, but
> only resources for which the server is authoritative.  Delivering Link
> header fields in a more timely fashion is more flexible and it allows
> clients to learn about resources they might like to load more quickly.
> """
> 
> Or something like that.

Yes probably something like this would be better.

Willy

Received on Wednesday, 2 August 2017 05:15:53 UTC